Research ArticleEvaluation of Dental Pulp Stem Cell Heterogeneity andBehaviour in 3D Type I Collagen Gels

Amr Alraies,1 Rachel J. Waddington,1 Alastair J. Sloan,2 and Ryan Moseley 1

1Regenerative Biology Group, Oral and Biomedical Sciences, School of Dentistry, Cardiff Institute of Tissue Engineering andRepair (CITER), College of Biomedical and Life Sciences, Cardiff University, Cardiff, UK2Melbourne Dental School, Faculty of Medicine, Dentistry and Health Sciences, The University of Melbourne, Melbourne, Australia

Correspondence should be addressed to Ryan Moseley; moseleyr@cardiff.ac.uk

Received 18 May 2020; Revised 27 August 2020; Accepted 3 September 2020; Published 10 September 2020

Academic Editor: Arianna Scuteri

Copyright © 2020 Amr Alraies et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Dental pulp stem cells (DPSCs) are increasingly being advocated for regenerative medicine-based therapies. However, significantheterogeneity in the genotypic/phenotypic properties of DPSC subpopulations exist, influencing their therapeutic potentials. Asmost studies have established DPSC heterogeneity using 2D culture approaches, we investigated whether heterogeneous DPSCproliferative and contraction/remodelling capabilities were further evident within 3D type I collagen gels in vitro. DPSCsubpopulations were isolated from human third molars and identified as high/low proliferative and multipotent/unipotent,following in vitro culture expansion and population doubling (PD) analysis. High proliferative/multipotent DPSCs, such as A3(30 PDs and 80 PDs), and low proliferative/unipotent DPSCs, such as A1 (17 PDs), were cultured in collagen gels for 12 days,either attached or detached from the surrounding culture plastic. Collagen architecture and high proliferative/multipotent DPSCmorphologies were visualised by Scanning Electron Microscopy and FITC-phalloidin/Fluorescence Microscopy. DPSCproliferation (cell counts), contraction (% diameter reductions), and remodelling (MMP-2/MMP-9 gelatin zymography) ofcollagen gels were also evaluated. Unexpectedly, no proliferation differences existed between DPSCs, A3 (30 PDs) and A1 (17PDs), although A3 (80 PDs) responses were significantly reduced. Despite rapid detached collagen gel contraction with A3 (30PDs), similar contraction rates were determined with A1 (17 PDs), although A3 (80 PDs) contraction was significantlyimpaired. Gel contraction correlated to distinct gelatinase profiles. A3 (30 PDs) possessed superior MMP-9 and comparableMMP-2 activities to A1 (17 PDs), whereas A3 (80 PDs) had significantly reduced MMP-2/MMP-9. Highproliferative/multipotent DPSCs, A3 (30 PDs), further exhibited fibroblast-like morphologies becoming polygonal withinattached gels, whilst losing cytoskeletal organization and fibroblastic morphologies in detached gels. This study demonstratesthat heterogeneity exists in the gel contraction and MMP expression/activity capabilities of DPSCs, potentially reflectingdifferences in their abilities to degrade biomaterial scaffolds and regulate cellular functions in 3D environments and theirregenerative properties overall. Thus, such findings enhance our understanding of the molecular and phenotypic characteristicsassociated with high proliferative/multipotent DPSCs.

1. Introduction

Adult human dental pulp stem cells (DPSCs) are increasinglybeing characterised and evaluated as a viable mesenchymalstem cell (MSC) source in the development of effective regen-erative medicine-based therapies for clinical use [1–3]. Suchconclusions are based on DPSCs being readily accessible

from the permanent dentition, in addition to their self-renewal, clonogenicity, and multilineage (e.g., osteogenic,chondrogenic, adipogenic, myogenic, and neurogenic) differ-entiation capabilities, comparable to those established forbone marrow-derived MSCs [4–6]. Consequently, DPSCshave already been shown to be beneficial to repair followingtransplantation into various animal model defects in vivo,

HindawiBioMed Research InternationalVolume 2020, Article ID 3034727, 12 pageshttps://doi.org/10.1155/2020/3034727

related to diseases and traumas within clinical fields such asDentistry, Orthopaedics, Neurology, Ophthalmology, andCardiology [1–3].

In order to replenish MSC populations and facilitate denovo tissue repair, exogenously sourced MSCs are invariablytransplanted into wound sites seeded within three-dimensional (3D) biomaterial scaffolds or hydrogels, whichact as cell carriers and provide mechanical support. Further-more, the plethora of synthetic, semisynthetic, and naturallyderived scaffold materials currently used endeavor to recapit-ulate the physiochemical properties of the extracellularmatrix (ECM) comprising the native stem cell niche micro-environment, which actively controls MSC self-renewal, pro-liferative, migratory, and differentiation responses [7–10].Scaffold-based 3D assemblies provide a spatial arrangementthat permit multiple cellular interactions with MSCs viaintegrin-based focal adhesions and Rho GTPases, which reg-ulate cytoskeletal organization and cellular morphology, inaddition to mediating cell signalling, mechanical force trans-duction, and adhesion to the biomaterial substratum [11, 12].Most natural and semisynthetic biomaterial scaffolds aredesigned to undergo gradual remodelling and degradationby matrix metalloproteinases (MMPs), such as gelatinasesMMP-2 andMMP-9, to be replaced by the newly synthesizedtissue whilst still providing a functional role in supportingMSC activity [13–16].

Despite such advances, fundamental challenges stillremain with the development of DPSC-based therapies forclinical applications, due to the significant heterogeneitybetween DPSCs isolated from dental pulp tissues, with indi-vidual subpopulations demonstrating major differences inproliferation and differentiation capabilities [4, 5, 17, 18].Consequently, despite heterogeneous DPSC populationsachieving >120 population doublings (PDs) in vitro, only20% of purified DPSCs are capable of proliferating >20PDs. Such features have significant implications for DPSCexploitation, as a major limitation of MSC-based therapiesis the extensive in vitro expansion required to produce suffi-cient cell numbers for clinical use, which eventually leads toproliferative decline, replicative senescence, and impairedMSC regenerative properties [19, 20].

Recent studies by ourselves have demonstrated thatwhilst high proliferative DPSCs are capable of >80 PDs inculture, low proliferating DPSCs only completed <40 PDsbefore senescence, correlating with DPSCs with high prolif-erative capacities possessing longer telomeres than less pro-liferative subpopulations [21, 22]. Low proliferative DPSCsenescence was also associated with the loss of stem cellmarker characteristics and impaired osteogenic/chondro-genic differentiation, favouring adipogenesis. In contrast,high proliferative DPSCs retained multipotency capabilities,only demonstrating impaired differentiation followingprolonged in vitro expansion (>60 PDs).

Therefore, due to the established genotypic and pheno-typic heterogeneity previously demonstrated to exist betweenhigh proliferative/multipotent and low proliferative/unipo-tent DPSC subpopulations utilising 2D monolayer cultures,this study investigated whether similar DPSC heterogeneityexisted in terms of their respective proliferation and abilities

to remodel/contract 3D type I collagen gels in vitro. Specifi-cally, the proliferative and ECM contraction/remodellingcapabilities of a high proliferative/multipotent DPSC sub-population, A3, were evaluated at early (nonsenescent) stages(at 30 PDs) and later (approaching senescence) stages (at 80PDs) in its proliferative lifespan and compared against a lowproliferative/unipotent DPSC, A1, approaching senescence(at 17 PDs). Type I collagen gels were employed, as they areextensively used in the development of natural biomaterialscaffolds for regenerative medicine applications [23, 24].

2. Materials and Methods

2.1. Dental Pulp Stem Cell Isolation and Characterisation ofProliferation/Differentiation Capabilities in 2D MonolayerCulture.Human DPSCs were isolated from third molar teethcollected from patients (all female, age 18-30 years) undergo-ing orthodontic extractions at the School of Dentistry, Car-diff University, UK. Teeth were collected in accordancewith the Declaration of Helsinki (2013), with informedpatient consent and ethical approval by the South East WalesResearch Ethics Committee of the National Research EthicsService (NRES), UK.

Single cell suspensions of dental pulp tissues wereobtained, with DPSCs preferentially selected and isolatedfrom cell suspensions by differential fibronectin adhesionassay [21]. Isolated cells were confirmed as DPSCs throughcell surface marker expression (positive for MSC markers,CD73, CD90, and CD105, negative for hematopoietic stemcell marker, CD45). DPSCs subsequently underwentextended culture expansion and characterisation as beinghigh or low proliferative and multi- or unipotent. Prolifera-tion analysis was based on the PDs reached by each DPSCsubpopulation prior to senescence, confirmed by the detec-tion of other senescence-related markers, including reducedtelomere lengths, positive senescence-associated β-galactosi-dase staining, and increased p53, p21waf1, and p16INK4a

expression [21].Differentiation analyses were based on the abilities of

each DPSC subpopulation to undergo osteogenic, chondro-genic, and adipogenic differentiation, via detection of estab-lished differentiation markers as previously described [21].Individual DPSC subpopulations were subsequently con-firmed as being high proliferative/multipotent DPSCs (suchas A3, capable of >80 PDs) or low proliferative/unipotentDPSCs (such as A1, capable of <40 PDs).

2.2. Preparation and Culture of 3D Type I Collagen Gels.DPSCs were seeded in type I collagen gels, based on themethod of Stephens et al. [25]. Type I rat tail collagen(>2mg/mL, First Link, Wolverhampton, UK) was reconsti-tuted in an α-modified Minimum Essential Medium(αMEM), containing ribonucleosides and deoxyribonucleo-sides, supplemented with 4mM L-glutamine, 100U/mL pen-icillin G sodium, 0.1μg/mL streptomycin sulphate,0.25μg/mL amphotericin (all ThermoFisher Scientific, Pais-ley, UK), and 100μM L-ascorbate 2-phosphate (Sigma,Poole, UK), to provide 1mg/mL solutions. Collagen solu-tions were neutralised by the dropwise addition of 5M

2 BioMed Research International

sodium hydroxide. Subsequently, aliquots (1mL) of the col-lagen solutions were added into each well of 12-well plateand maintained at 37°C/5% CO2 for 45min. High proliferati-ve/multipotent DPSC subpopulations, A3, at early (nonse-nescent, 30 PDs) and late (approaching senescence, 80 PDs)stages in their proliferative lifespans, and low proliferative/u-nipotent DPSCs, A1 (approaching senescence, 17 PDs), weresuspended in αMEM, containing 20% gelatinase-free, foetalcalf serum (FCS, ThermoFisher Scientific, prepared fromcomplete FCS according to [26]). DPSCs (5 × 104 cells/mL)in gelatinase-free αMEMwere subsequently added to the col-lagen gels. Collagen gels (n = 3 per DPSC subpopulation)were maintained at 37°C/5% CO2 for 12 days, with gelsremaining either attached to the surrounding tissue cultureplastic or detached at Day 0 using a sterile filter tip. Acellulartype I collagen gels were also established as controls.

2.3. Scanning Electron Microscopy (SEM) Analysis ofAcellular Collagen Gels. The architecture of type I collagenfibres within the acellular gels was assessed by Scanning Elec-tron Microscopy (SEM). Acellular collagen gels were estab-lished as described above and fixed with 2.5%glutaraldehyde (Agar Scientific, Stansted, UK) in 0.1M caco-dylate buffer for 2 h at room temperature. Gels were subse-quently washed (×3) with 0.1M cacodylate buffer anddehydrated through a graded ethanol series: 30%, 50%,70%, and 100% for 15min each, and gels allowed to dry over-night. Dried hydrogels were sputter-coated and SEM imagestaken at multiple locations throughout each gel, using aS4800 Scanning Electron Microscope (Hitachi, Tokyo,Japan), operated at 10 kV.

2.4. DPSC Proliferation in Type I Collagen Gels. Collagen gelswere also established as described above, for the high prolif-erative/multipotent DPSC subpopulation, A3, at early (non-senescent, 30 PDs) and late (approaching senescence, 80PDs) stages in their proliferative lifespans, and low prolifera-tive/unipotent DPSCs, A1 (approaching senescence, 17 PDs),in order to assess their respective proliferative capabilitiesover 12 days in culture. DPSCs were recovered at Days 1, 2,4, 6, and 12 from collagen gels by enzymatic digestion, as pre-viously described [25]. At each time point, collagen gels weresolubilized with collagenase A (400μL, 2mg/mL, Roche,West Sussex, UK), at 37°C/5% CO2 for 2 h. This was followedby incubation with 0.05% trypsin/0.53mM EDTA (100μL,ThermoFisher Scientific), at 37°C/5% CO2 for 20min. DPSCswere recovered by centrifugation (1,500 g/5min) and viablecell numbers determined using a Neubauer haemocytometer,with 0.02% Trypan Blue solution (Sigma). Each experimentwas performed on 3 separate occasions.

2.5. Type I Collagen Gel Contraction by DPSCs. DPSC-seededcollagen gels were established as described above. The rela-tive extents of remodelling and contraction of the type I col-lagen gels by high proliferative/multipotent DPSCsubpopulation, A3, at early (nonsenescent, 30 PDs) and late(approaching senescence, 80 PDs) stages in their proliferativelifespans, and low proliferative/unipotent DPSCs, A1(approaching senescence, 17 PDs), were quantified from 3

separate gel diameter measurements performed on each rep-licate sample (n = 3 per DPSC subpopulation), at Days 1, 2, 4,6, and 12 in the culture. Average contraction values obtainedwere expressed as % reduction in gel diameter, compared togel diameters at 0 h. Conditioned medium was also collectedfrom each individual well for the analysis of relative gelati-nase (MMP-2 and MMP-9) activities at the same time pointsas the gel diameter measurements. Each experiment wasperformed on 3 separate occasions.

2.6. Assessment of Gelatinase (MMP-2 and MMP-9)Activities. The relative amounts of pro- and active MMP-2and MMP-9 activities produced by high proliferative/multi-potent DPSC subpopulations, A3, at early (nonsenescent,30 PDs) and late (approaching senescence, 80 PDs) stagesin their proliferative lifespans, and low proliferative/unipo-tent DPSCs, A1 (approaching senescence, 17 PDs), in type Icollagen gels, were determined by gelatin zymography basedon the method of Stephens et al. [27]. Collected conditionedmedia (25μL) were mixed with equal volumes of nonreduc-ing Laemmli loading buffer and separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE), on precast10% gelatin zymography gels (Ready Gel 10% Gelatin Zymo-gram Gels, Bio-Rad Laboratories, Hemel Hempstead, UK),using a Mini-PROTEAN® Tetra Cell System (Bio-Rad Labo-ratories) at 20mA. SDS was removed from each SDS-PAGEgel by washing in 2.5% Triton X-100 solution (Sigma), for1 h at room temperature. Following electrophoresis, gelati-nases within the SDS-PAGE gels were activated by incuba-tion in 25mM Tris-HCl buffer, pH7.6, containing 5mMcalcium chloride (Sigma), 25mM sodium chloride (Thermo-Fisher Scientific), and 5% Brij 35 (Sigma), at 37°C overnight.Gels were stained with 0.05% Coomassie Blue solution(Sigma) in 12% acetic acid and 54% methanol (both Thermo-Fisher Scientific) and destained in 7.5% acetic acid/5% metha-nol solution. Gel images were captured, with MMP-2 andMMP-9 identified by the appearance of clear bands at compa-rable molecular weights to loaded MMP-2 and MMP-9 stan-dards, respectively. The relative amounts of MMP-2 andMMP-9 in captured images were determined by densitometry,using ImageJ® Software (https://imagej.nih.gov/ij/). Eachexperiment was performed on 3 separate occasions.

2.7. SEM and Fluorescence Microscopy Analysis of DPSCMorphologies in Collagen Gels. The cellular morphologies ofthe nonsenescent (30 PDs), high proliferative/multipotentDPSC subpopulation, A3, in type I collagen gels, were exam-ined by both SEM and via actin cytoskeleton staining usingFITC-phalloidin and Fluorescence Microscopy. DPSC-seeded collagen gels were established, as described above.For SEM, collagen gels were fixed, processed, and visualised,as per the methods detailed above. For Fluorescence Micros-copy, DPSC-seeded collagen gels were washed withphosphate-buffered saline (PBS, 1 × 1mL) and fixed with4% paraformaldehyde (1mL, Santa Cruz, Dallas, USA) for10min at room temperature. Gels were washed with PBS(2 × 1mL) at 5min interval and treated with 0.3% TritonX-100 (1mL) for 30min at room temperature, under con-stant agitation. Gels were washed with Tris-buffered saline

3BioMed Research International

(TBS, pH7.5, 2 × 1mL) and blocked with 1% bovine serumalbumin (BSA, Fraction V, ThermoFisher Scientific) in TBS(1% BSA-TBS) for 1 h at room temperature, under constantagitation. DPSC-seeded gels were stained with FITC-phalloidin (20μg/mL, 1mL, Sigma) in 1% BSA-TBS andincubated for 1 h at 4°C, under darkness and constant agita-tion. Gels were washed with TBS (2 × 1mL) at 5min intervaland images captured by Fluorescence Microscopy (OlympusProvis AX70 Microscope, Olympus UK Ltd., Southend-on-Sea, UK).

2.8. Statistical Analysis. Each experiment was performed onn = 3 independent occasions. Data were expressed as mean± standard error of mean (SEM). Graphical data were ini-tially confirmed as exhibiting homoscedasticity usingLevene’s test. Having shown that the averages for all experi-mental group data were considered equal, graphical datawere subsequently compared statistically by Analysis of Var-iance (ANOVA), performed using GraphPad InStat 3(GraphPad Software Inc., La Jolla, USA). Statistical signifi-cance was considered at p < 0:05.

3. Results

3.1. SEM Analysis of Acellular Collagen Gels. RepresentativeSEM images of the acellular collagen gels established usingthe protocols described above are shown in Figures 1(a)and 1(b). The matrix architecture is typical for type I collagengels of this nature, with the formation of a dense fibrillarynetwork of narrow, elongated, and crosslinked collagen fibres[28, 29]. Consequently, the collagen fibre arrangements pro-duced resulted in relatively small pore sizes being formedwithin the collagen gels overall.

3.2. DPSC Proliferation in Type I Collagen Gels. Mean cellnumbers obtained for high proliferative/multipotent DPSCsubpopulations, A3, at early (nonsenescent, 30 PDs) and late(approaching senescence, 80 PDs) stages in their proliferativelifespans, and low proliferative/unipotent DPSCs, A1(approaching senescence, 17 PDs), in attached and detachedtype I collagen gels over 12 days in culture, are shown inFigure 2. For attached collagen gels, all DPSC subpopulationsexhibited similar cell number profiles for the first 4 days(Figure 2(a)). However, A3 (30 PDs) and A1 (17 PDs) subse-quently began to increase in cell number to Day 6, followedby reductions in cell numbers by Day 12 (all p > 0:05 betweenA3, 30 PDs, and A1, 17 PDs). In contrast, A3 (80 PDs) failedto exhibit any obvious increases in cell number beyond Day 4(p < 0:05 versus A3, 30 PDs, and A1, 17 PDs, at Day 6).

With detached collagen gels, DPSCs A3 (30 PDs) and A1(17 PDs) did not show any apparent increases in cell numberduring the 12 days in culture, unlike the scenario underattached type I collagen gel conditions (all p > 0:05,Figure 2(b)). However, A3 (80 PDs) demonstrated compara-ble cell number profiles under detached collagen gel condi-tions, to their attached gel counterparts over the 12-dayculture period (all p > 0:05).

3.3. Type I Collagen Gel Contraction by DPSCs. Representa-tive type I collagen gel contraction images for nonsenescent

(30 PDs), high proliferative/multipotent DPSC subpopula-tions, A3, over 12 days in culture are shown in Figure 3(a).Comparisons of mean collagen contraction values, expressedas the % reduction in gel diameter compared to Day 0, dem-onstrated that all DPSC subpopulations maintained inattached collagen gels over the 12-day culture period didnot exhibit any gel contraction due to collagen gelattachment to the adjacent tissue culture plastic (all p > 0:05;data not shown). In contrast, DPSC subpopulations indetached type I collagen gels with high proliferative/multipo-tent DPSC subpopulations, A3, at early (nonsenescent, 30PDs) and late (approaching senescence, 80 PDs) stages intheir proliferative lifespans, and low proliferative/unipotentDPSCs, A1 (approaching senescence, 17 PDs), showed vari-able contraction rates over time (Figure 3(b)). For both A3(30 PDs) and A1 (17 PDs), gels displayed gradual decreasesin diameter/contraction over the initial 4 days in culture,although gels seeded with A3 (30 PDs) displayed greatercontraction, compared to A1 (17 PDs). However, from Day6 onwards, gels containing A1 (17 PDs) exhibited equivalentdecreases in diameter/contraction to A3 (30 PDs) gels. Con-sequently, no significant differences were identified in therates of type I collagen contraction between gels seeded withA3 (30 PDs) and A1 (17 PDs) throughout the 12-day cul-ture period (all p > 0:05). In contrast to A3 (30 PDs) andA1 (17 PDs), gels seeded with A3 (80 PDs) displayedalmost no shrinkage over the first 6 days in culture, withonly a total reduction in diameter of ≅26% from its originalsize by Day 12 (Figure 3(b)). As such, gels containingDPSC subpopulation A3 (80 PDs) demonstrated signifi-cantly delayed gel contraction (all p < 0:001‐0:01), com-pared to those seeded with A3 (30 PDs) and A1 (17 PDs).

3.4. Gelatinase Activities in DPSC-Seeded Collagen Gels. Rep-resentative MMP-2 and MMP-9 gelatin zymography imagesfor high proliferative/multipotent DPSC subpopulations, A3,at early (nonsenescent, 30 PDs) and late (approaching senes-cence, 80 PDs) stages in their proliferative lifespans, and lowproliferative/unipotent DPSCs, A1 (approaching senescence,17 PDs), in detached type I collagen gels over 12 days in cul-ture, are shown in Figure 4. MMP zymography demonstratedthat pro-MMP-2 was detectable for all DPSC subpopulationsanalysed, being detectable from Day 1 onwards, whilst activeMMP-2 was only detectable with A3 (30 PDs) and A1 (17PDs, Figure 4(a)). Although MMP-2 activities graduallyincreased over time in culture, A3 (80 PDs) demonstratedthe lowest levels of detectable pro- and active MMP-2 overall(Figure 4(b)), being significantly less than A3 (30 PDs) andA1 (17 PDs) at all time points (p < 0:001‐0:01 for pro-MMP-2). However, significant differences in pro-MMP-2activities between A3 (30 PDs) and A1 (17 PDs) were onlydetermined at Day 6 (p < 0:001). In contrast to MMP-2, A3(30 PDs) was the only DPSC subpopulation to exhibit detect-able pro-MMP-9 activities, being significantly higher thanA1 (17 PDs, all p < 0:001) and A3 (80 PDs, all p < 0:001)throughout the culture duration (Figures 4(a) and 4(c)).

3.5. DPSC Morphologies in Type I Collagen Gels. As a conse-quence of the rapid detached collagen gel contraction and

4 BioMed Research International

unique MMP-9 expression/activity properties of the nonse-nescent (30 PDs), high proliferative/multipotent DPSC sub-population, A3, the cell morphologies of this subpopulationwere further examined in type I collagen gels by SEM andFluorescence Microscopy. SEM analysis demonstrated thatDPSCs possessed good biocompatibility with the collagengels overall, with viable cells detectable throughout the 12-day culture duration. DPSCs were shown to attach to colla-gen fibres and begin to spread at Day 1, with DPSCs largelydisplaying flattened, spindle-shaped, or polygonal morphol-ogies overall (Figures 5(a) and 5(b)). At Day 12, DPSCs hadspread further resulting in predominantly larger and moreirregular 3D cuboidal or polygonal morphologies, due tosuperior interactions with the collagen matrices comparedto Day 1, through numerous contacts between cell extensionsand collagen fibres and the propagation of shortpseudopodia-like protrusions into the ECM from the cellmembranes (Figures 5(c) and 5(d)).

Representative Fluorescence Microscopy images of highproliferative/multipotent DPSC subpopulation, A3 (30PDs), cytoskeletal properties and morphologies in attached

and detached type I collagen gels at Day 1 are shown inFigure 6. DPSCs exhibited limited morphological differencesthroughout culture overall. However, whereas DPSCs inattached type I collagen gels presented very clear actin stressfibres with a fibroblastic-like appearance (Figure 6(a)),DPSCs maintained in detached collagen gels appeared to losetheir fibroblastic appearance with less prominent actin fibres(Figure 6(b)).

4. Discussion

This study aimed to determine whether the well-establishedvariations in the proliferative, differentiation and other stemcell properties between different DPSC subpopulations in the2Dmonolayer culture were further identifiable in 3D cultureswithin type I collagen gels. The concept of DPSC heterogene-ity is well-established, with the presence of DPSC subpopula-tions in dental pulp tissues possessing contrastingproliferative and differentiation capabilities [4, 5, 17, 18].Furthermore, recent corroborating findings have reportedkey differences between DPSCs in the relative susceptibilities

(a) (b)

Figure 1: SEM images of the acellular type I collagen gels established using the protocols described herein, at (a) ×4,000 and (b) ×15,000magnifications.

500000

400000

300000

200000

100000

Mea

n D

PSC

coun

ts (a

ttach

ed g

els)

00 1 2 4

Time (days)

A3 (30 PDs)A3 (80 PDs)A1 (17 PDs)

6 12

⁎ ⁎

(a)

Mea

n D

PSC

coun

ts (d

etac

hed

gels)

0 1 2 4Time (days)

6 12

200000

100000

0

A3 (30 PDs)A3 (80 PDs)A1 (17 PDs)

(b)

Figure 2: Mean cell numbers for high proliferative/multipotent DPSC subpopulation, A3, at early (nonsenescent, 30 PDs) and late(approaching senescence, 80 PDs) stages in their proliferative lifespans, and low proliferative/unipotent DPSCs, A1 (approachingsenescence, 17 PDs), following seeding in (a) attached and (b) detached type I collagen gels over 12 days in culture. N = 3, mean ± SEM, ∗

p < 0:05.

5BioMed Research International

to replicative senescence, correlating with contrasting telo-mere lengths and the differentiation capabilities of individualsubpopulations [21, 22]. However, despite differences in thegenotypic/phenotypic characteristics of individual DPSCsubpopulations in vitro, to date, no studies had directlyconfirmed whether similar differences in responses existwhen such DPSCs are seeded within 3D type I collagen gels.Therefore, the present study compared the proliferative andECM contraction/remodelling capabilities of high proliferati-ve/multipotent DPSC subpopulations, such as A3, at early(nonsenescent) stages (at 30 PDs) and later (approachingsenescence) stages (at 80 PDs) in its proliferative lifespan,versus low proliferative/unipotent DPSCs, such as A1, whilstapproaching senescence (at 17 PDs).

A great deal of evidence exists confirming that cellsrespond differently when cultured within 3D gel constructs,in comparison to the 2D monolayer culture, ascribed to thesimilarities between the 3D gel environment and the naturalECM comprising the MSC niche [8]. Furthermore, the statusof 3D type I collagen and other gel types, in terms of whetherthese are detached or remain attached to the surrounding tis-sue culture plastic, has been previously shown to regulate cel-lular behaviour, including morphology, proliferation, anddifferentiation, in addition to contraction itself [30–34].SEM and Fluorescence Microscopy analyses demonstratedthat DPSCs, such as the high proliferative/multipotent sub-population, A3, at nonsenescent stages in its proliferative life-span (at 30 PDs), readily interacted and spread within type I

collagen gels, particularly possessing spindle-shaped,fibroblast-like morphologies with obvious actin stress fibres,during early culture in attached type I collagen gels. Morepolygonal DPSC morphologies developed during prolongedculture due to superior interactions with the type I collagenmatrices. In contrast, this DPSC subpopulation lost itsfibroblast-like morphologies with less prominent actin fibresevident, when seeded within detached type I collagen gels.Collagen gel contraction by resident cells causes mechanicalforces to be induced within the surrounding ECM, capableof influencing cellular alignment depending on the forcesencountered [30]. As described below, type I collagen gelcontraction by this DPSC subpopulation was particularlyapparent during early culture stages, such as between Day 0and Day 1. Thus, it was hypothesized that any prominent dif-ferences in the cytoskeletal organization between theattached and detached type I collagen gels would be particu-larly evident by Day 1, as was proven to be the case. Hence, asevident here, cells within attached gels alone develop isomet-ric tension by forming prominent cytoskeletal actin stressfibres and focal adhesions with the surrounding ECM, per-mitting pseudopodia to aid their attachment or movementwithin the type I collagen environment [35–37].

Various cell types have also previously been shown toundergo mechanical stress-induced proliferation withinattached collagen gels [31–33]. Despite increased DPSCnumbers towards the latter stages of culture within attachedtype I collagen gels, relatively modest increases in DPSC

Day 1 Day 2 Day 4 Day 6 Day 12

(a)

120

100

80

60

40

20

% ty

pe I

colla

gen

gel c

ontr

actio

n

00 1 2 4 6

Time (days)12

A3 (30 PDs)A3 (80 PDs)A1 (17 PDs)

⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎ ⁎⁎⁎

⁎⁎

(b)

Figure 3: (a) Images of detached type I collagen gel contraction by nonsenescent (30 PDs), high proliferative/multipotent DPSCsubpopulation, A3, over 12 days in culture. (b) Mean % detached type I collagen gel contraction by high proliferative/multipotent DPSCsubpopulation, A3, at early (nonsenescent, 30 PDs) and late (approaching senescence, 80 PDs) stages in their proliferative lifespans, andlow proliferative/unipotent DPSCs, A1 (approaching senescence, 17 PDs), over 12 days in culture. N = 3, mean ± SEM, ∗∗∗p < 0:001, ∗∗p <0:01.

6 BioMed Research International

proliferation were evident overall throughout the 12-day cul-ture period. In contrast, detached gels showed minimal pro-liferation for all DPSC subpopulations analysed [38, 39].However, cell number comparisons between high proliferati-ve/multipotent DPSC subpopulations, A3, at early (nonse-

nescent, 30 PDs) and late (approaching senescence, 80 PDs)stages in their proliferative lifespans, and low proliferative/u-nipotent DPSCs, A1 (approaching senescence, 17 PDs),within attached type I collagen gels, revealed significantlyimpaired A3 (80 PDs) proliferative responses in 3D culture,

A3 (30 PDs)Time (days)

Pro-MMP-9

Pro-MMP-2Active MMP-2

Pro-MMP-2Active MMP-2

Pro-MMP-2Active MMP-2

92 kDa1 2 4 6 12

1 2 4 6 12

1 2 4 6 12

72 kDa62 kDa

Time (days)

72 kDa62 kDa

Time (days)

72 kDa62 kDa

A3 (80 PDs)

A1 (17 PDs)

(a)

1 2 4Time (days)

6 120

5000

Pro-

MM

P-2

band

inte

nsity

(OD

× m

m2 )

10000

15000

20000

25000

⁎⁎ ⁎⁎⁎

⁎⁎⁎ ⁎⁎⁎

⁎⁎⁎ ⁎⁎⁎

⁎⁎⁎

⁎⁎⁎

⁎⁎⁎⁎⁎⁎ ⁎⁎

A3 (30 PDs)A3 (80 PDs)A1 (17 PDs)

(b)

A3 (30 PDs)A3 (80 PDs)A1 (17 PDs)

10

5000

Pro-

MM

P-9

band

inte

nsity

(OD

× m

m2 )

10000

15000

20000

25000

2 4Time (days)

6 12

⁎⁎⁎ ⁎⁎⁎⁎⁎⁎ ⁎⁎⁎⁎⁎⁎ ⁎⁎⁎⁎⁎⁎ ⁎⁎⁎

⁎⁎⁎ ⁎⁎⁎

(c)

Figure 4: (a) Representative gelatin zymography gel images of pro- and active MMP-9/MMP-2 activities for high proliferative/multipotentDPSC subpopulation, A3, at early (nonsenescent, 30 PDs) and late (approaching senescence, 80 PDs) stages in their proliferative lifespans,and low proliferative/unipotent DPSCs, A1 (approaching senescence, 17 PDs), following seeding in detached type I collagen gels over 12days in culture. Pro-MMP-2, active MMP-2, and pro-MMP-9 were detectable at 72 kDa, 62 kDa, and 92 kDa, respectively. Densitometryquantification of (b) pro-MMP-2 and (c) pro-MMP-9 activities for high proliferative/multipotent DPSC subpopulation, A3 (30 PDs and80 PDs), and low proliferative/unipotent DPSC subpopulation (A1, 17 PDs), following seeding in detached type I collagen gels over 12days in culture. N = 3, mean ± SEM, ∗∗∗p < 0:001, ∗∗p < 0:01.

7BioMed Research International

compared to earlier in their proliferative lifespan (30 PDs).Intriguingly, no significant differences in proliferation wereidentifiable between A3 (30 PDs) and A1 (17 PDs), despiteconsiderable differences in their proliferative capacities andsusceptibilities to replicative senescence being previouslyidentified in the 2D monolayer culture (>80 PDs versus<40 PDs, respectively; [21]). Therefore, obvious differencesin proliferative capacities between A3 (80 PDs) and A3 (30PDs) may be accounted for by the relative stages in their pro-liferative lifespan where this DPSC subpopulation wasassessed, although the precise reasons behind such discrep-ancies in the relative proliferative responses of A3 (30 PDs)

and A1 (17 PDs) between 2D and 3D environments remainto be elucidated.

Although studies have determined that cells proliferatepoorly in detached gels due to their growth arrest withinthe G0/G1 phase of the cell cycle [40], the principle reasonsunderlying the maintenance of cellular proliferative capabili-ties within attached gels remain to be fully determined.Nonetheless, cell signalling pathways, such as ERK or MAPkinases, are acknowledged to respond to various stimuli,including mechanical stress, whilst such signalling cascadesare not initiated in detached collagen gels, due to the absenceof isometric tension and cytoskeletal actin stress fibre

(a) (b)

(c) (d)

Figure 5: SEM images of nonsenescent (30 PDs), high proliferative/multipotent DPSC subpopulation, A3, following seeding in type Icollagen gels at (a) Day 1 (×700 magnification), (b) Day 1 (×2,500 magnification), (c) Day 12 (×1,100 magnification), and (d) Day 12(×3,500 magnification).

500 𝜇m

(a)

500 𝜇m

(b)

Figure 6: Fluorescence Microscopy images of FITC-phalloidin cytoskeletal staining of nonsenescent (30 PDs), high proliferative/multipotentDPSC subpopulation, A3, following seeding in (a) attached type collagen gels and (b) detached type collagen gels, at Day 1. ×100magnification. Scale bar = 500 μm.

8 BioMed Research International

formation [41, 42]. Therefore, it is plausible that similar dif-ferences in ERK and MAPK signalling exist between theDPSC subpopulations within attached and detached type Icollagen gels herein and warrant further investigation.

Cellular functions within the stem cell niche or biomate-rial scaffolds are tightly regulated by the ECMmicroenviron-ment, remodelled by proteinases such as MMPs [43, 44]. Ofthe different MMP groups known, gelatinases MMP-2 andMMP-9 have common roles in mediating MSC remodellingof the niche and biomaterial scaffold degradation, therebyfacilitating cellular activities [13–16]. Analysis of the relativeabilities of high proliferative/multipotent DPSC subpopula-tions, A3, at early (nonsenescent, 30 PDs) and late(approaching senescence, 80 PDs) stages in their proliferativelifespans, and low proliferative/unipotent DPSCs, A1(approaching senescence, 17 PDs), to remodel and contractdetached type I collagen gels, demonstrated initial rapid col-lagen gel contraction with A3 (30 PDs), although comparablerates of gel contraction between A3 (30 PDs) and A1 (17PDs) were identified overall during the 12-day culture dura-tion. In contrast, A3 (80 PDs) displayed significantlyimpaired gel contraction capabilities, as a likely consequenceof this being at the latter stages in its proliferative lifespanand more senescent than its lesser senescent A3 (30 PDs)counterparts. Such gel contraction findings were ascribed tocontrasting MMP-2 and MMP-9 activities detected for eachDPSC subpopulation, with significantly elevated levels ofpro- and active forms of MMP-2 and MMP-9 being respon-sible for collagen gel contraction by a high proliferative/mul-tipotent DPSC subpopulation, A3 (30 PDs), and pro- andactive forms of MMP-2 alone for gel contraction by a lowproliferative/unipotent DPSC subpopulation, A1 (17 PDs).However, A3 (80 PDs) demonstrated significant reducedpro- and active MMP-2 and MMP-9 activities overall.

Similarly impaired type I collagen gel reorganizationalcapabilities have previously been reported with other senes-cent cell populations, partly as a consequence of reducedMMP-2 expression and activities [45, 46]. Replicative senes-cence is also well-established to induce significant alterationsinMSC genotype and phenotype, leading to impaired cellularregenerative properties and signalling mechanisms via thesecretome associated with the senescence-associated secre-tory phenotype (SASP) [20, 47]. However, despite the SASPbeing accompanied by excessive expression of proinflamma-tory mediators, including MMPs, excessive MMP-2 and/orMMP-9 activities by more senescent A3 (80 PDs) were notdetectable herein, as DPSC subpopulation, A3, lost its abili-ties to express MMP-2 and MMP-9 with extensive cultureexpansion in line with previous findings [45, 46]. Further-more, in addition to limited type I collagen gel contractionand MMP-2/MMP-9 activities by A3 (80 PDs), superiorremodelling efficiencies were only evident with A3 (30 PDs)at certain time points, compared to the more senescent sub-population, A1 (17 PDs).

Although there are many similarities between gelatinases,MMP-2 and MMP-9, in terms of their structures and sub-strates, differences in their cellular sources, regulation, andactivation do exist [43, 44]. Gelatinase expression and activa-tion are regulated by numerous factors, including cytokines

and growth factors. Mechanical stress is also a key factorcapable of regulating MMP expression and activities within3D gel environments during contraction, whilst MSC func-tions have also been shown to be highly regulated by MMPs[14, 16, 48–51]. Despite the inactive zymogens of MMP-2and MMP-9 existing as proforms due to the presence oflatent propeptide domains, differences in prodomain cleav-age and proteolytic activation by other MMPs or proteinasesare acknowledged. Pro-MMP-2 activation can be mediatedvia a broad range of MMPs, most significantly through com-plex formation with membrane-type MMPs (MT-MMPs)and tissue inhibitor of metalloproteinases (TIMPs) [43, 44].However, unlike pro-MMP-2, MT-MMP-/TIMP-dependentmechanisms of activation have not been described for pro-MMP-9, which is mediated by alternative MMP-drivenmechanisms. Indeed, although MMP-2 is a key contributorto the proteolytic activation of pro-MMP-9, it is establishedthat pro-MMP-9 predominantly remains in its latent pro-form during the culture of many different cell types, even inthe presence of active MMP-2 [52]. Thus, a paradox remainsas to the lack of active MMP-9 under cell culture conditions,although structural and catalytic differences between thesegelatinases and their sequestration within the ECM havebeen suggested to contribute to this phenomenon [43, 44].Nonetheless, the nonsenescent high proliferative/multipo-tent DPSC subpopulation, A3 (30 PDs), was unique in dem-onstrating pro-MMP-9 detection. Thus, as pro-MMP-9 isrecognised to possess significant catalytic activity, even withthe propeptide domain intact [53], the absence of activeMMP-9 should not curtail the potential significance of pro-MMP-9 in facilitating type I collagen gel remodelling bythese DPSC subpopulations. Indeed, this mechanism couldbe important for MSC maintenance in a quiescent statewithin the stem cell niche or in enhancing the 3D biomaterialscaffold remodelling/degradation capabilities of this DPSCsubpopulation [15, 49]. Furthermore, pro-MMP-9 detectionin the high proliferative/multipotent subpopulation, A3, atnonsenescent stages in its proliferative lifespan (at 30 PDs),may indicate another feature of the heterogeneous naturebetween DPSCs isolated from dental pulp tissues [4, 5, 17,18, 21, 22]. However, we can only speculate at present as tothe reasons underlying such differences in type I collagengel contraction and MMP remodelling capabilities betweenhigh proliferative/multipotent and low proliferative/unipo-tent DPSCs, as patient donor features, developmental origins,stem cell niche sources, and/or the stages of commitment ofthese DPSC subpopulations would undoubtedly be potentialinfluences on these cellular characteristics.

5. Conclusions

As with previous studies highlighting the phenotypic andgenotypic heterogeneity which exists between high prolifera-tive/multipotent and low proliferative/unipotent DPSC sub-populations utilising 2D monolayer cultures, this studydemonstrates that heterogeneity is also evident in terms ofthe gel contraction capabilities and MMP expression/activityprofiles of individual DPSC subpopulations, following seed-ing within 3D type I collagen gels. Despite distinct differences

9BioMed Research International

in proliferative capabilities and susceptibilities to replicativesenescence between DPSC subpopulations previously beingestablished during the extended 2D monolayer culture, con-trasting proliferative responses were unexpectedly less appar-ent within 3D collagen gels. Such variations in type I collagenremodelling mechanisms by MMP-2 and MMP-9 betweenhigh proliferative/multipotent and low proliferative/unipo-tent DPSC subpopulations may reflect novel differences intheir respective abilities to degrade biomaterial scaffoldsand regulate cellular functions in 3D environments followingin vivo transplantation, thereby influencing their individualregenerative properties overall. Consequently, these findingsalso aid our understanding of the molecular and phenotypicproperties associated with high proliferative/multipotentDPSCs, which may be exploited for their selective screeningand isolation from dental pulp tissues for regenerative medi-cine applications.

Data Availability

The data used to support the findings of this study are avail-able from the corresponding author upon request.

Conflicts of Interest

There are no conflicts of interest from any author.

Authors’ Contributions

RM, AJS, and RJW conceived and designed the experiments.AA performed the experiments. AA analysed the data. RMwrote the paper. RJW, AJS, and AA contributed to the paperreview.

Acknowledgments

This work was supported by a PhD studentship awarded toDr. Amr Alraies by Albawani Company, Saudi Arabia. Theauthors also wish to thank Dr. Nafiseh Badiei (College ofEngineering, Swansea University, UK), for her help in under-taking the Scanning Electron Microscopy (SEM) analysis ofthe acellular and DPSC-seeded type I collagen gels.

References

[1] E. Ledesma-Martínez, V. M. Mendoza-Núñez, andE. Santiago-Osorio, “Mesenchymal stem cells derived fromdental pulp: a review,” Stem Cells International, vol. 2016, Arti-cle ID 4709572, 12 pages, 2016.

[2] E. P. Chalisserry, S. Y. Nam, S. H. Park, and S. Anil, “Thera-peutic potential of dental stem cells,” Journal of Tissue Engi-neering, vol. 8, 17 pages, 2016.

[3] E. Anitua, M. Troya, andM. Zalduendo, “Progress in the use ofdental pulp stem cells in regenerative medicine,” Cytotherapy,vol. 20, no. 4, pp. 479–498, 2018.

[4] S. Gronthos, J. Brahim, W. Li et al., “Stem cell properties ofhuman dental pulp stem cells,” Journal of Dental Research,vol. 81, no. 8, pp. 531–535, 2016.

[5] A. J. Sloan and R. J. Waddington, “Dental pulp stem cells:what, where, how?,” International Journal of Paediatric Den-tistry, vol. 19, no. 1, pp. 61–70, 2009.

[6] N. Nuti, C. Corallo, B. M. F. Chan, M. Ferrari, and B. Gerami-Naini, “Multipotent differentiation of human dental pulp stemcells: a literature review,” Stem Cell Reviews and Reports,vol. 12, no. 5, pp. 511–523, 2016.

[7] N. Zippel, M. Schulze, and E. Tobiasch, “Biomaterials andmesenchymal stem cells for regenerative medicine,” RecentPatents on Biotechnology, vol. 4, no. 1, pp. 1–22, 2010.

[8] H. Donnelly, M. Salmeron-Sanchez, and M. J. Dalby, “Design-ing stem cell niches for differentiation and self-renewal,” Jour-nal of the Royal Society Interface, vol. 15, no. 145, article20180388, 2018.

[9] V. V. Hiew, S. F. B. Simat, and P. L. Teoh, “The advancementof biomaterials in regulating stem cell fate,” Stem Cell Reviewsand Reports, vol. 14, no. 1, pp. 43–57, 2018.

[10] Z. Liu, M. Tang, J. Zhao, R. Chai, and J. Kang, “Looking intothe future: toward advanced 3D biomaterials for Stem-Cell-Based regenerative medicine,” Advanced Materials, vol. 30,no. 17, article e1705388, 2018.

[11] S. I. Fraley, Y. Feng, R. Krishnamurthy et al., “A distinctive rolefor focal adhesion proteins in three-dimensional cell motility,”Nature Cell Biology, vol. 12, no. 6, pp. 598–604, 2010.

[12] S. Martino, F. D'Angelo, I. Armentano, J. M. Kenny, andA. Orlacchio, “Stem cell-biomaterial interactions for regenera-tive medicine,” Biotechnology Advances, vol. 30, no. 1, pp. 338–351, 2012.

[13] J. Kim, I. S. Kim, T. H. Cho et al., “In vivo evaluation of MMPsensitive high-molecular weight HA-based hydrogels for bonetissue engineering,” Journal of Biomedical Materials ResearchPart A, vol. 95, no. 3, pp. 673–681, 2010.

[14] R. K. Schneider, A. Puellen, R. Kramann et al., “The osteogenicdifferentiation of adult bone marrow and perinatal umbilicalmesenchymal stem cells and matrix remodelling in three-dimensional collagen scaffolds,” Biomaterials, vol. 31, no. 3,pp. 467–480, 2010.

[15] H. Zhang, L. Zhou, and W. Zhang, “Control of scaffold degra-dation in tissue engineering: a review,” Tissue Engineering PartB: Reviews, vol. 20, no. 5, pp. 492–502, 2014.

[16] M. Daviran, H. S. Caram, and K. M. Schultz, “Role of cell-mediated enzymatic degradation and cytoskeletal tensionon dynamic changes in the rheology of the pericellularregion prior to human mesenchymal stem cell motility,”ACS Biomaterials Science and Engineering, vol. 4, no. 2,pp. 468–472, 2017.

[17] S. Gronthos, M. Mankani, J. Brahim, P. G. Robey, and S. Shi,“Postnatal human dental pulp stem cells (DPSCs) in vitroand in vivo,” Proceedings of the National Academy of Sciencesof the United States of America, vol. 97, no. 25, pp. 13625–13630, 2000.

[18] G. T.-J. Huang, S. Gronthos, and S. Shi, “Mesenchymal stemcells derived from dental tissues vs. those from other sources:their biology and role in regenerative medicine,” Journal ofDental Research, vol. 88, no. 9, pp. 792–806, 2009.

[19] J. Campisi and F. d'Adda di Fagagna, “Cellular senescence:when bad things happen to good cells,” Nature Reviews Molec-ular Cell Biology, vol. 8, no. 9, pp. 729–740, 2007.

[20] Y. Li, Q. Wu, Y. Wang, L. Li, H. Bu, and J. Bao, “Senescence ofmesenchymal stem cells (review),” International Journal ofMolecular Medicine, vol. 39, no. 4, pp. 775–782, 2017.

10 BioMed Research International

[21] A. Alraies, N. Y. A. Alaidaroos, R. J. Waddington, R. Moseley,and A. J. Sloan, “Variation in human dental pulp stem cell age-ing profiles reflect contrasting proliferative and regenerativecapabilities,” BMC Cell Biology, vol. 18, no. 1, p. 12, 2017.

[22] A. Alraies, E. Canetta, R. J. Waddington, R. Moseley, and A. J.Sloan, “Discrimination of dental pulp stem cell regenerativeheterogeneity by single-cell Raman spectroscopy,” Tissue Engi-neering Part C: Methods, vol. 25, no. 8, pp. 489–499, 2019.

[23] B. D. Walters and J. P. Stegemann, “Strategies for directing thestructure and function of three-dimensional collagen biomate-rials across length scales,” Acta Biomaterialia, vol. 10, no. 4,pp. 1488–1501, 2014.

[24] S. R. Chowdhury, M. F. Mh Busra, Y. Lokanathan et al., “Col-lagen type I: a versatile biomaterial,” Advances in ExperimentalMedicine Biology, vol. 1077, pp. 389–414, 2018.

[25] P. Stephens, K. J. Davies, T. al-Khateeb, J. P. Shepherd, andD. W. Thomas, “A comparison of the ability of intra-oraland extra-oral fibroblasts to stimulate extracellular matrixreorganization in a model of wound contraction,” Journal ofDental Research, vol. 75, no. 6, pp. 1358–1364, 2016.

[26] H. S. Azzam and E. W. Thompson, “Collagen-induced activa-tion of theMr 72,000 type IV collagenase in normal and malig-nant human fibroblastoid cells,” Cancer Research, vol. 52,no. 16, pp. 4540–4544, 1992.

[27] P. Stephens, P. Grenard, P. Aeschlimann et al., “Crosslinkingand G-protein functions of transglutaminase 2 contribute dif-ferentially to fibroblast wound healing responses,” Journal ofCell Science, vol. 117, no. 15, pp. 3389–3403, 2004.

[28] T. Schuetz, N. Richmond, M. E. Harmon, J. Schuetz,L. Castaneda, and K. Slowinska, “The microstructure of colla-gen type I gel cross-linked with gold nanoparticles,” Colloidsand Surfaces B: Biointerfaces, vol. 101, pp. 118–125, 2013.

[29] O. Moreno-Arotzena, J. Meier, C. del Amo, and J. García-Aznar, “Characterization of fibrin and collagen gels for engi-neering wound healing models,” Materials, vol. 8, no. 4,pp. 1636–1651, 2015.

[30] M. Eastwood, V. C. Mudera, D. A. McGrouther, and R. A.Brown, “Effect of precise mechanical loading on fibroblastpopulated collagen lattices: morphological changes,” CellMotility and Cytoskeleton, vol. 40, no. 1, pp. 13–21, 1998.

[31] K. Yamamoto, T. Sokabe, T. Watabe et al., “Fluid shear stressinduces differentiation of Flk-1-positive embryonic stem cellsinto vascular endothelial cells in vitro,” American Journal ofPhysiology: Heart and Circulatory Physiology, vol. 288, no. 4,pp. H1915–H1924, 2005.

[32] J. H.-C. Wang and B. P. Thampatty, “Chapter 7 Mechanobiol-ogy of adult and stem cells,” International Review of Cell andMolecular Biology, vol. 271, pp. 301–346, 2008.

[33] A. J. F. Stops, K. B. Heraty, M. Browne, F. J. O'Brien, and P. E.McHugh, “A prediction of cell differentiation and proliferationwithin a collagen–glycosaminoglycan scaffold subjected tomechanical strain and perfusive fluid flow,” Journal of Biome-chanics, vol. 43, no. 4, pp. 618–626, 2010.

[34] A. J. Steward, D. R. Wagner, and D. J. Kelly, “The pericellularenvironment regulates cytoskeletal development and the dif-ferentiation of mesenchymal stem cells and determines theirresponse to hydrostatic pressure,” European Cells and Mate-rials, vol. 25, pp. 167–178, 2013.

[35] K. A. Beningo, M. Dembo, and Y. L. Wang, “Responses offibroblasts to anchorage of dorsal extracellular matrix recep-tors,” Proceedings of the National Academy of Sciences of the

United States of America, vol. 101, no. 52, pp. 18024–18029,2004.

[36] H. Jiang and F. Grinnell, “Cell–matrix entanglement andmechanical anchorage of fibroblasts in three-dimensional col-lagen matrices,” Molecular Biology of the Cell, vol. 16, no. 11,pp. 5070–5076, 2005.

[37] J. S. Gabbay, J. B. Heller, S. A. Mitchell et al., “Osteogenicpotentiation of human adipose-derived stem cells in a 3-dimensional matrix,” Annals of Plastic Surgery, vol. 57, no. 1,pp. 89–93, 2006.

[38] H. Rosenfeldt and F. Grinnell, “Fibroblast quiescence and thedisruption of ERK signaling in mechanically unloaded colla-gen matrices,” Journal of Biological Chemistry, vol. 275, no. 5,pp. 3088–3092, 2000.

[39] E. Hadjipanayi, V. Mudera, and R. A. Brown, “Close depen-dence of fibroblast proliferation on collagen scaffold matrixstiffness,” Journal of Tissue Engineering and Regenerative Med-icine, vol. 3, no. 2, pp. 77–84, 2009.

[40] T. Kono, T. Tanii, M. Furukawa et al., “Cell cycle analysis ofhuman dermal fibroblasts cultured on or in hydrated type Icollagen lattices,” Archives of Dermatological Research,vol. 282, no. 4, pp. 258–262, 1990.

[41] C. Ruwhof and A. van der Laarse, “Mechanical stress-inducedcardiac hypertrophy: mechanisms and signal transductionpathways,” Cardiovascular Research, vol. 47, no. 1, pp. 23–37,2000.

[42] J. Fringer and F. Grinnell, “Fibroblast quiescence in floating orreleased collagen matrices,” Journal of Biological Chemistry,vol. 276, no. 33, pp. 31047–31052, 2001.

[43] T. Klein and R. Bischoff, “Physiology and pathophysiology ofmatrix metalloproteases,” Amino Acids, vol. 41, no. 2,pp. 271–290, 2011.

[44] V. Vargová, M. Pytliak, and V. Mechírová, “Matrix metallo-proteinases,” EXS, vol. 103, pp. 1–33, 2012.

[45] H. Cook, P. Stephens, K. J. Davies, D. W. Thomas, and K. G.Harding, “Defective extracellular matrix reorganization bychronic wound fibroblasts is associated with alterations inTIMP-1, TIMP-2, and MMP-2 activity,” Journal of Investiga-tive Dermatology, vol. 115, no. 2, pp. 225–233, 2000.

[46] S. Enoch, M. A. Peake, I. Wall et al., “'Young' oral fibroblastsare geno/phenotypically distinct,” Journal of Dental Research,vol. 89, no. 12, pp. 1407–1413, 2010.

[47] N. Malaquin, A. Martinez, and F. Rodier, “Keeping the senes-cence secretome under control: molecular reins on thesenescence-associated secretory phenotype,” ExperimentalGerontology, vol. 82, pp. 39–49, 2016.

[48] D. Karamichos, R. A. Brown, and V. Mudera, “Collagen stiff-ness regulates cellular contraction and matrix remodeling geneexpression,” Journal of Biomedical Materials Research Part A,vol. 83A, no. 3, pp. 887–894, 2007.

[49] G. Kasper, J. D. Glaeser, S. Geissler et al., “Matrix metallopro-tease activity is an essential link between mechanical stimulusand mesenchymal stem cell behavior,” Stem Cells, vol. 25,no. 8, pp. 1985–1994, 2007.

[50] K. Bott, Z. Upton, K. Schrobback et al., “The effect of matrixcharacteristics on fibroblast proliferation in 3D gels,” Biomate-rials, vol. 31, no. 32, pp. 8454–8464, 2010.

[51] E. W. Howard, B. J. Crider, D. L. Updike et al., “MMP-2expression by fibroblasts is suppressed by the myofibroblastphenotype,” Experimental Cell Research, vol. 318, no. 13,pp. 1542–1553, 2012.

11BioMed Research International

[52] M. Toth, I. Chvyrkova, M. M. Bernardo, S. Hernandez-Bar-rantes, and R. Fridman, “Pro-MMP-9 activation by the MT1-MMP/MMP-2 axis and MMP-3: role of TIMP-2 and plasmamembranes,” Biochemical and Biophysical Research Commu-nications, vol. 308, no. 2, pp. 386–395, 2003.

[53] M. Björklund and E. Koivunen, “Gelatinase-mediated migra-tion and invasion of cancer cells,” Biochimica et BiophysicaActa, vol. 1755, no. 1, pp. 37–69, 2005.

12 BioMed Research International